Blog & News
Recent developments, analysis, and updates from the world of cybersecurity and AI safety.
Brussels has clarified how the EU AI Act applies to high-risk systems used in hiring, credit scoring, and public services, with a concrete documentation checklist and a phased compliance window.
A mainstream browser now automatically flags images carrying C2PA provenance data, surfacing an AI-generated badge without requiring an extension or any technical know-how from the user.
NIST has issued new guidance on vetting third-party AI models and training data, treating a poisoned model the same way mature security teams already treat a compromised software dependency.
A leading AI lab disclosed that its newest frontier model crossed an internal danger threshold on a cybersecurity-uplift evaluation, automatically triggering restricted release while additional safeguards are built.
A survey of application security teams finds AI coding assistants reproducing a distinct, recurring set of flaws, and doing it identically across many unrelated codebases at once.
Vishing attacks using AI-cloned executive voices are rising, with attackers needing only a short public recording to produce a convincing impersonation for a wire-transfer request.
DeepMind has open-sourced a testing framework that automates prompt-injection red-teaming, giving smaller teams access to a class of security testing previously limited to well-resourced AI labs.
Chatbot safety filters get bypassed constantly, not through hacking, but through clever phrasing. Here is the structural reason that keeps happening.
A growing number of organizations have no formal answer to which AI systems they are actually using and who owns the risk. Here is what building that answer from zero tends to look like.